Why the gap years are their own problem. Almost every other retirement decision assumes Medicare is waiting at 65. But roughly half of people leave full-time work before then, and the day the paychecks stop is usually the day employer health coverage stops too. From that moment until your 65th birthday you are on your own for insurance — and skipping coverage isn't a real option, because a single unlucky hospitalization can run into six figures. This is the piece early retirees most often underestimate: Chapter 10 covered healthcare costs broadly; this chapter is specifically about who pays the premiums between your last day of work and Medicare.
Option 1 — COBRA (the expensive bridge). Federal COBRA lets you keep your former employer's exact plan for up to 18 months after you leave. The coverage is identical and your doctors stay in-network — but you now pay the entire premium, both your old share and the part your employer used to cover, plus a 2% administrative fee. What felt like a $200/month payroll deduction can suddenly be $700–$1,800/month. COBRA is best as a short stopgap: to finish a course of treatment, keep a specific doctor through year-end, or simply cover you until an ACA plan starts. For a multi-year bridge it's usually the costliest route, because it comes with no income-based subsidy.
Option 2 — a spouse's plan, or part-time work with benefits. Before anything else, check whether a still-working spouse can add you to their employer plan (leaving a job is a "qualifying life event" that opens a special enrollment window), or whether a part-time or "bridge" job offers coverage — some employers extend benefits to part-timers, and it can be worth working a few hours a week purely for the insurance. These are often cheaper and simpler than anything you buy on your own.
Option 3 — the ACA marketplace (the workhorse). For most early retirees the answer is an individual plan from the Affordable Care Act marketplace — Healthcare.gov or a state-run exchange. Coverage is guaranteed-issue: insurers can't deny you or charge you more for pre-existing conditions, and premiums vary only by age, location, plan tier, and tobacco use. Plans come in metal tiers — Bronze (low premium, high deductible), Silver, Gold, Platinum — and you enroll during open enrollment (roughly Nov 1–Jan 15) or a special enrollment period triggered by losing job coverage. On paper these plans look expensive; the reason they're viable is the premium tax credit.
How the premium tax credit actually works. The PTC doesn't knock a flat amount off your bill. Instead it says: you should pay no more than a certain percentage of your income for a "benchmark" plan, and the government pays the rest. The mechanics: 1. Find the benchmark. The credit is pegged to the second-lowest-cost Silver plan (SLCSP) available to you — not necessarily the plan you pick. 2. Compute your expected contribution. Based on your household income as a percentage of the federal poverty level (FPL), a sliding scale sets the share of income you're expected to pay. 3. The credit is the gap. Subsidy = (benchmark Silver premium) − (your expected contribution). That dollar amount can then be applied to any metal-tier plan you choose. So if the benchmark Silver plan costs $1,400/month and your income means you should pay $500/month, your credit is $900/month — and you could spend it on a cheaper Bronze plan (paying little or nothing) or put it toward a richer Gold plan.
It's your MAGI that drives everything. The income figure the marketplace uses is a modified adjusted gross income (MAGI): your AGI plus a few add-backs — tax-exempt muni interest, the untaxed portion of Social Security, and foreign-earned income. Crucially, it's a household number and it's for the coverage year, so you estimate it in advance and the subsidy is advanced to your insurer monthly (the "advance premium tax credit," APTC). This is exactly why the credit is such a powerful lever for early retirees: unlike a wage earner, you largely choose your MAGI by deciding which buckets to spend from (Chapter 20). Money pulled from a taxable account is mostly return of principal plus a little gain; Roth withdrawals don't count as income at all; only traditional IRA/401(k) withdrawals, conversions, interest, dividends, and realized gains push MAGI up. An early retiree with a big taxable and Roth balance can live comfortably while showing a modest MAGI — and collect a large subsidy.
The sliding scale, the floor, and the cliff. Two boundaries define the subsidy zone, and both matter: - The Medicaid floor (~138% FPL in expansion states). If your MAGI is too low, you don't get a marketplace subsidy — you get routed to Medicaid instead (below roughly 138% of poverty in the 40+ states that expanded it). That's free or near-free coverage, but with narrower provider networks; some early retirees deliberately keep income above the Medicaid line to stay on a marketplace plan. In the handful of non-expansion states, falling below 100% FPL creates a genuine "coverage gap" with no help at all. - The 400% FPL "subsidy cliff" — live again in 2026. Under the ACA's original design, subsidies vanished entirely once income crossed 400% of poverty — one dollar over and you lost the whole credit. The American Rescue Plan (2021) and Inflation Reduction Act temporarily removed that cliff and capped the benchmark premium at 8.5% of income. Those enhancements expired on December 31, 2025, and the hard cliff returned for the 2026 plan year: clear $62,600 (single) or $84,600 (a couple) by one dollar and the whole credit is gone. This is not a rounding error — an unsubsidized benchmark Silver plan for a 60-year-old averages about $15,900 a year in 2026, and KFF estimates a 60-year-old earning $65,000 now pays roughly $10,400 more per year than under the enhanced rules. For an early retiree this is the single most expensive line in the plan, and it turns MAGI management from a nice-to-have into the main event. Congress may restore some version of the enhancement; until it does, treat 400% as a wall, and confirm which regime is in effect for your coverage year before planning around an income target.
Cost-sharing reductions — the Silver bonus. If your income is under 250% of poverty, choosing a Silver plan unlocks cost-sharing reductions (CSRs): the insurer quietly lowers your deductible, copays, and out-of-pocket maximum, often turning a mediocre Silver plan into something closer to Gold or Platinum coverage for the same premium. Below ~200% FPL the effect is dramatic. The trap is picking Bronze for its low premium and forfeiting CSRs you were entitled to — under 250% FPL, Silver is usually the smart tier.
The Roth-conversion tension. Here's the sophisticated trade-off unique to these years. The pre-Social-Security, pre-RMD gap is the classic window for cheap Roth conversions (Chapter 13) and for realizing gains in the 0% capital-gains bracket (Chapter 11) — both of which raise MAGI. But higher MAGI shrinks your ACA subsidy. So a dollar of Roth conversion in a bridge year can carry a hidden cost: not just the income tax, but the lost premium credit on top — an effective marginal rate that can quietly exceed 30–40% once the subsidy phase-out is counted. There's no universal answer. Some retirees prioritize the subsidy (keep MAGI low until 65, convert aggressively afterward); others decide the long-term RMD and tax-torpedo relief from converting now outweighs a few years of forgone subsidy. The right call depends on the size of your pre-tax balance, your benchmark premium, and how close you are to the floor or cliff — which is exactly the kind of thing worth modeling rather than guessing.
Reconcile at tax time (Form 8962). Because your subsidy is based on estimated income, you square up when you file: Form 8962 compares the advance credit you received against the credit your actual MAGI entitled you to. Estimate too low and you'll repay some or all of the excess; estimate too high and you get the difference back as a refund. For early retirees whose income comes from lumpy withdrawals and conversions, this makes year-end income management essential — a surprise December capital-gain distribution or an over-aggressive conversion can trigger a clawback, or in a cliff year, wipe out the entire subsidy.
Managing MAGI for a subsidy
A couple retires at 61 with $1.4M split across taxable, traditional, and Roth accounts, needing about $80,000/year to live and coverage until Medicare at 65. Left unmanaged — say they fund the year mostly from the traditional IRA — their MAGI lands near $90,000 — above the $84,600 that is 400% of poverty for a couple. Under the 2026 rules that's over the cliff, full stop: their benchmark Silver premium of ~$2,200/month (older couples pay the highest age-rated premiums) gets no credit — roughly $26,000 out of pocket. Now they restructure: they spend mostly from the taxable account (largely basis) and Roth, pulling just enough from the traditional IRA to hold MAGI around $70,000 (~330% FPL). Their expected contribution falls to a set share of income, and the subsidy now covers a large chunk of that $2,200 premium — plausibly cutting their net cost by $10,000–$18,000 a year. Same lifestyle, same $80,000 of spending; the only thing that changed was which accounts the money came from. Over four bridge years that's tens of thousands of dollars — the payoff for treating MAGI as a dial you control.